Data Exfiltration via MCP Server Prompt Injection Targeting Cursor
Attack on AI1 Research / proof of conceptPrompt injectionAgent hijackData leak
Researchers showed that a malicious Model Context Protocol server can deliver an indirect prompt injection to the Cursor IDE. This injection tricked the AI agent into executing shell commands to exfiltrate configuration files.
Details
- type
- Attack on AIAn AI model, agent or AI-powered product was the thing exploited, abused or leaked.
- categories
- Prompt injectionprimaryUntrusted content (web pages, documents, emails, tool output) steers a model into ignoring its instructions.
- Agent hijackAn AI agent is tricked into unauthorized actions: tool misuse, data exfiltration through the agent, or acting against its principal.
- Data leakTraining data, chat logs, system prompts, model weights or other AI-system data exposed.
- Prompt injectionprimary
- owasp llm top 10
- mitre atlas
- Both rows are derived from the type and categories (OWASP 2025 edition, ATLAS release 2026.09), not classified per record. How
- severity
- 1 · Research / proof of conceptDemonstrated by researchers; no known real-world exploitation.
- targets
- Framework / tooling, Enterprise deployment
- products
- Cursor, Model Context Protocol
- incident date
- 24 Jun 2025
- first reported
- 30 Jan 2026
- origin
- MITRE ATLASDerived from MITRE ATLAS (Apache 2.0, © MITRE).
- classifier confidence
- 0.95 · published at or above 0.80; see the methodology
- tags
- mcp · prompt_injection · cursor · data_leak
- record
- created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z · json
Sources · 2
- 2026-01-30Data Exfiltration via an MCP Server used by CursorMITRE ATLAS · atlas.mitre.org/studies/AML.CS0045
- 2026-01-30Simulating a Context-Poisoning Vulnerable MCP Server: A POCbackslash.security · backslash.security/blog/simulating-a-vulnerable-mcp-server-for-context-poisoning
Titles and links belong to their publishers. The summary above is written by GCD AI RADAR in its own words; data is CC BY-SA 4.0.
Related incidents
- 2026-03-31Credential Exfiltration via Remote Poisoned MCP Tool1 Researchon AIPrompt inj.Agent hijack2 src
- 2026-06-29Cross-Site Scripting via Prompt Manipulation in Lenovo AI Chatbot1 Researchon AIPrompt inj.Agent hijack2 src
- 2025-11-07AIKatz Extracts Authentication Tokens from LLM Desktop Apps1 Researchon AIData leakAgent hijack2 src
- 2025-11-07Data Exfiltration from Slack AI via Indirect Prompt Injection1 Researchon AIPrompt inj.Agent hijack2 src
- 2025-11-07Data Exfiltration via Prompt Injections in Copilot Studio Agent1 Researchon AIPrompt inj.Agent hijack3 src
- 2026-07-31Claude Code GitHub Action Secret Exposure via Prompt Injection1 Researchon AIPrompt inj.Agent hijack2 src